Recovery And Restoration Of The Seagrass Halodule Wrightii After Boat Propeller Scar Damage In A Pole-troll Zone In Mosquito Lag

Transcription

1 University of Centrl Florid Electronic Theses nd Disserttions Msters Thesis (Open Access) Recovery And Restortion Of The Segrss Hlodule Wrightii After Bot Propeller Scr Dmge In A Pole-troll Zone In Mosquito Lg 2008 Ktherine Grlow University of Centrl Florid Find similr works t: University of Centrl Florid Lirries Prt of the Biology Commons STARS Cittion Grlow, Ktherine, "Recovery And Restortion Of The Segrss Hlodule Wrightii After Bot Propeller Scr Dmge In A Pole-troll Zone In Mosquito Lg" (2008). Electronic Theses nd Disserttions This Msters Thesis (Open Access) is rought to you for free nd open ccess y STARS. It hs een ccepted for inclusion in Electronic Theses nd Disserttions y n uthorized dministrtor of STARS. For more informtion, plese contct lee.dotson@ucf.edu.

2 RECOVERY AND RESTORATION OF THE SEAGRASS HALODULE WRIGHTII AFTER BOAT PROPELLER SCAR DAMAGE IN A POLE-TROLL ZONE IN MOSQUITO LAGOON, FLORIDA y KATHERINE RUTH GRABLOW B.S. University of Centrl Florid, 2006 A thesis sumitted in prtil fulfillment of the requirements for the degree of Mster of Science in the Deprtment of Biology in the College of Sciences t the University of Centrl Florid Orlndo, FL Fll 2008

3 ABSTRACT This study comined documenttion of four ot propeller scr types in Hlodule wrightii segrss eds in Mosquito Lgoon, Florid with mnipultive field experiments to document scr recovery times with nd without restortion. Scr types rnged from the most severe scr type (Type 1) with trench formtion which hd no roots or shoots in the trench, to the lest severe (Type 4) scrs tht hd no depth, intct roots nd shoots shorter thn the surrounding cnopy. For 110 mesured existing scrs, the frequency of ech scr type ws 56% for Type 1, 10% for Type 2, 7% for Type 3, nd 27% for Type 4. In the first mnipultive experiment, experimentl scrs were creted to document the nturl recovery time of H. wrightii for ech scr severity within one yer. Type 4 scrs recovered to the control shoot density t 2 months, while Types 1, 2, nd 3 scrs did not fully recover in one yer. Men estimted recovery for H. wrightii is expected in 25 months for Type 1, nd 19 months for Types 2 nd 3. For the second mnipultive experiment, three restortion methods were tested on Type 1 scrs over 1 yer period. Restortion methods included: (1) plnting H. wrightii in the scr trench, (2) filling the trench with snd, nd (3) filling with snd plus plnting H. wrightii. There ws complete mortlity of ll trnsplnts t 2 months nd only 25% of scrs retined fill snd fter 1 yer. With dense djcent segrss eds, nturl recovery ws more successful thn ny of my restortion ttempts. Thus, I suggest tht mngers should concentrte on preventing segrss destruction rther thn restortion. ii

4 To my husnd Seth, nd my fmily for ll their help, love, nd support. iii

5 ACKNOWLEDGMENTS Thnk you Seth Grlow, Louis Brown, Molly Brown, Nicole Mrtucci, Juli Leissing, Andre Brer, Json Ledgrd, Hyley Ashworth, Iris Howell, Mtthew Mitchell, Lurie Hollidy, Christy Akers, Shnnon Hckett, Kristen Kneifl, Ethn Nsh, Elizeth Bourss, Gisel Hrper, Justin Bridges, Rchel Odom, Wei Yun, Ssh Brodsky, the Student Conservtion Assocition interns, nd Cnverl Ntionl Seshore Physicl Lor Crew for helping set up nd mesure the experiment in the field. Thnk you Lind Wlters, Dvid Jenkins, Roert Virnstein, Troy Rice, Roert Dy, Lori Morris, Luren Hll, John Stiner, nd Mike Legre, for your guidnce, support nd encourgement! Thnk you to the University of Centrl Florid Grdute Studies Reserch nd Mentoring Fellowship, Deprtment of Biology, Indin River Lgoon Ntionl Estury Progrm, St. Johns River Wter Mngement District, Grden Clu of Americ, nd Interntionl Women s Fishing Assocition for supporting nd funding this project. iv

6 TABLE OF CONTENTS CHAPTER 1: GENERAL INTRODUCTION... 1 CHAPTER 2: BASELINE SCAR MEASUREMENTS... 6 METHODS... 6 Study orgnism... 6 Study re... 7 Mesurements of existing propeller scrs... 8 Intr-scr mesurements... 9 Sediment nlyses... 9 Sttisticl nlyses RESULTS Scr severity ctegories Generl scr mesurements Intr-scr mesurements Scr sediment nlysis DISCUSSION CHAPTER 3: MANIPULATIVE EXPERIMENTS METHODS Study re Lyout nd monitoring of experiments Nturl recovery of segrss in experimentl propeller scrs Restortion of experimentl propeller scrs Sediment core nlysis RESULTS Site consistency Nturl recovery experiment Sesonl vritions in nturl recovery Restortion experiment Erosion oservtions DISCUSSION Nturl recovery Segrss restortion Mngement recommendtions Restortion versus prevention Future of segrss in Mosquito Lgoon APPENDIX A: LIST OF FIGURES APPENDIX B: LIST OF TABLES LITERATURE CITED v

7 LIST OF FIGURES Figure 1: Mp of experimentl scr loctions in the Pole-Troll Zone of Mosquito Lgoon, Florid. Boundry of the Pole-Troll Zone is mrked y dotted line with lck circles. Chnnels include: the Atlntic Intercostl Wterwy (ICW) s solid lck line, nd the inner Pole-Troll Zone chnnels re dotted lines within the Pole-Troll Zone oundries Figure 2: Hlodule wrightii segrss ed in Mosquito Lgoon, FL Figure 3: Loction of 110 propeller scrs mesured in Mosquito Lgoon, FL. Scr type is color coded to show the frequency of ech scr type. From this view, the entire rnge of mesured scrs is demonstrted, however not ll scrs cn e seen due to overlp Figure 4: Cnonicl Discriminnt Anlysis clssifiction of 110 propeller scrs in Mosquito Lgoon. Scrs were sorted into four groups. Type 1 scrs re white circles, Type 2 scrs re lck squres, Type 3 scrs re lck tringles, nd Type 4 scrs re white dimonds Figure 5: Scr types identified from mesuring 110 scrs in Mosquito Lgoon. Aove digrm simplifies wht ws seen in ech of the photogrphs elow Figure 6: Frequency of scr types in Mosquito Lgoon (n = 110) Figure 7: Regression of Type 1 scr width versus scr depth. Scr Type 1 ws used to show the est correltion (r 2 = 0.476, p < 0.001), ecuse other scr types (2-4) did not hve depth. 43 Figure 8: Men percent of sediment grin size frctions nd orgnic content per scr type. Percent sediment frction sizes re listed in order < 1mm is light grey cross htch, 1 mm is drk grey cross htch, 2 mm is solid lck, 5 mm is solid white, > 5mm is solid light grey, nd percent orgnic content is solid drk grey. Using one-wy ANOVA, only scr Type 1 nd 4 were significntly different in ll frction sizes, nd size < 1mm showed the gretest difference Figure 9: Cnonicl Discriminnt Anlysis clssifiction of 110 propeller scrs sediment frctions nd distnce mesurements. Type 1 scrs re white circles, Type 2 scrs re lck squres, Type 3 scrs re lck tringles, nd Type 4 scrs re white dimonds. Sediment types did not show pttern in reltion to distnce Figure 10: Lyout of one lock of experimentl scrs. Ech lock ws 18 x18 m nd ws divided into 9 squres, of which only 8 hd scr tretment. A squre ws 6 x 6 m contining single scr tretment in the center Figure 11: Setup of one set of tretments in lrge grid (lock). An 18 x 18 m perimeter ws mrked off with yellow rope, nd PVC pole ws plced every 6 m. Scrs were creted in the center of ech 6 x 6 m squre nd mrked with short PVC pole t ech end. All PVC grid poles were mrked with GPS loction nd then removed. After the scrs were creted, only the 2 PVC scr mrkers remined to mrk the ends of ech scr vi

8 Figure 12: PVC stke mrking the end of n experimentl scr Figure 13: PVC plow 32 cm in dimeter. Device used to mke experimentl propeller scrs the sme width, depth, nd scr slope ngle Figure 14: Arrow points to H. wrightii ttched to grden stple with twist tie in scr filled with snd. This is n eroded stple, mking it is esier to see how H. wrightii ws ttched with red twist tie Figure 15: Men wter temperture on smpling dtes (± S.E.) t loction of mnipultive experiments (n = 3) Figure 16: Men slinity (± S.E.) t loction of mnipultive experiments on smpling dtes (n = 3) Figure 17: Men wter clrity depth (± S.E.) t loction of mnipultive experiments on smpling dtes (n = 3). Mximum mesurement possile using turitdity tue ws 60 cm Figure 18: Men wter depth (± S.E.) of experimentl scr loctions of mnipultive experiments (n = 10). There were no significnt differences of men wter depth mong replicte locks (Kruskl-Wllis rnk test p > 0.006) Figure 19: Men surrounding scr H. wrightii density (± S.E.) per 25 x 25 cm re in experimentl scr replicte locks. There were no significnt differences mong surrounding densities in replicte locks (Kruskl-Wllis rnk test p > 0.006) Figure 20: Men surrounding scr H. wrightii percent cover (± S.E.) per 25 x 25 cm re in experimentl scr replicte locks. There were no significnt differences mong surrounding H. wrightii men percent cover in replicte locks (Kruskl-Wllis rnk test: p > 0.006) Figure 21: Men surrounding scr H. wrightii cnopy height (± S.E.) in experimentl scr replicte locks. There were no significnt differences mong surrounding men surrounding cnopy height in replicte locks (Kruskl-Wllis rnk test: p > 0.006) Figure 22: Nturl recovery experimentl scr H. wrightii shoot density per 25 x 25 cm re (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, Type 4 re white dimonds, nd control re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 23: Nturl recovery experimentl scr H. wrightii percent cover per 25 x 25 cm re (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, Type 4 re white dimonds, nd control re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters vii

9 Figure 24: Nturl recovery experimentl scr H. wrightii cnopy height (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, Type 4 re white dimonds, nd control re lck tringles. Mesurements were nlyzed with Kruskl- Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 25: Nturl recovery experimentl scr percent root cover per 25 x 25 cm re (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, nd Type 4 re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 26: Nturl recovery experimentl scr percent lef litter cover per 25 x 25 cm re (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, nd Type 4 re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 27: Nturl recovery experimentl scr depth (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, nd Type 4 re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 28: Nturl recovery experimentl scr width (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, nd Type 4 re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 29: Nturl recovery experimentl scr slope ngle density (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, nd Type 4 re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 30: Nturl recovery experiment men percent sediment frctions (± S.E.) from efore scr cretion nd from within scrs fter 1 yer (Kruskl-Wllis: p < 0.05). Br sections re leled in order from left to right: initil, Type 1 (finl), Type 2 (finl), Type 3 (finl), Type 4 (finl), nd control (finl) Figure 31: Nturl recovery experiment men percent orgnic content (± S.E) from efore scr cretion nd from within scrs fter 1 yer (Kruskl-Wllis: p < 0.05) Figure 32: Restortion experimentl scr H. wrightii density shoot density per 25 x 25 cm (± S.E). Controls re white tringles, Type 1 re lck circles, plnt-only re lck squres, fill only re white dimonds, nd plnt nd fill re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < viii

10 Tretments tht were significntly different on single dy re shown with different letters Figure 33: Restortion experimentl scr H. wrightii percent cover per 25 x 25 cm re (± S.E). Control re white tringles, Type 1 re lck circles, plnt-only re lck squres, fill only re white dimonds, nd plnt nd fill re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 34: Restortion experimentl scr H. wrightii cnopy height (± S.E). Control re white tringles, Type 1 re lck circles, plnt-only re lck squres, fill only re white dimonds, nd plnt nd fill re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 35: Restortion experimentl scr percent root cover per 25 x 25 cm re (± S.E). Type 1 re lck circles, plnt-only re white tringles, fill only re lck squres, nd plnt nd fill re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 36: Restortion experimentl scr percent lef litter cover per 25 x 25 cm re (± S.E). Type 1 re lck circles, plnt-only re white tringles, fill only re lck squres, nd plnt nd fill re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 37: Restortion experimentl scr depth (± S.E). Control re white tringles, Type 1 re lck circles, plnt-only re lck squres, fill only re white dimonds, nd plnt nd fill re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 38: Restortion experimentl scr width (± S.E). Control re white tringles, Type 1 re lck circles, plnt-only re lck squres, fill only re white dimonds, nd plnt nd fill re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 39: Restortion experimentl scr slope ngle density (± S.E). Control re white tringles, Type 1 re lck circles, plnt-only re lck squres, fill only re white dimonds, nd plnt nd fill re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters Figure 40: Restortion experiment men percent sediment frctions (± S.E.) from efore scr cretion nd from within scrs fter 1 yer (Kruskl-Wllis: p < 0.05). Br sections re ix

11 leled in order from left to right: initil, control (finl), Type 1 (finl), plnt only (finl), fill only (finl), nd plnt nd fill (finl) Figure 41: Restortion experiment men percent orgnic content (± S.E) from efore scr cretion nd from within scrs fter 1 yer (Kruskl-Wllis: p < 0.05) Figure 42: Types of erosion oserved in experimentl scrs: A) holes, B) wshouts, nd C) snd ptches. The lck rectngle represents the originl scr dimensions, rown is re sediment, drk rown represents depth, nd the green represents segrss Figure 43: Bioturtion cused y stingry. Stingry creted hole nd unuried row of Hlodule wrightii trnsplnts on stples, then left ehind pile of feces. In the left picture the rrow points to close-up of excrement pile in hole, nd in right picture the rrow points to close up of the eroded stples next to the excrement pile Figure 44: Boter illeglly using min engine motor in Pole-Troll Zone nd running directly over the experiment site Figure 45: This PVC pole mrking n experimentl scr tht ws hit y min engine ot motor in less thn 0.5 m of wter x

12 LIST OF TABLES Tle 1: Previous studies documenting existing nd experimentl propeller scr dmge nd recovery of segrsses. Unville informtion ws denoted with dsh, nd recovery times tht included restortion method were denoted with n sterisk Tle 2: Cnonicl Discriminnt Anlysis results clssifying 110 propeller scr mesurements into four groups. Tle shows the percentge of scrs correctly clssified ccording to the most importnt predictor vriles: shoot density, scr depth, scr % root cover, nd scr cnopy height Tle 3: Men scr mesurements (± S.E.) for ech scr type from 110 scrs Tle 4: Men scr density mesurements (± S.E.) per 25 x 25 cm re for ech scr type from 110 scrs Tle 5: Cnonicl Discriminnt Anlysis results from clssifying sediment frctions of 110 scrs. This tle shows the percentge of scrs correctly clssified ccording to originl identifictions. Sediment frctions were not good predictor of scr types Tle 6: Men experimentl scr mesurements (± S.E.) for scr Types 1-4 nd control in nturl recovery experiment. Mesurements were tken eight times from August 2007 through August Tle 7: Kruskl-Wllis test for men differences in mesurements mong scr types in nturl recovery experiment. Since tests were repeted eight times (once for ech time period), the significnce level ws corrected to p = using the Bonferroni method Tle 8: Men experimentl scr mesurements (± S.E.) for three restortion tretments, nd Type 1 scr control in restortion experiment. Mesurements were tken eight times from August 2007 through August Tle 9: Survivl of three restortion tretments for Type 1 scrs etween August 2007 nd August Tle 10: Kruskl-Wllis test for men differences in mesurements mong three restortion tretments, the Type 1 scr control, nd the segrss control in restortion experiment. Since tests were repeted eight times (once for ech time period), the significnce level ws corrected to p = using the Bonferroni method Tle 11: Men (± S.E.) numer of scrs hving type of erosion per time period in ech tretment xi

13 CHAPTER 1: GENERAL INTRODUCTION Segrss ecosystems re one of the most productive res on erth, lthough they cover only % of the ocen glolly (Durte 2002, Duffy 2006, Erftemeijer nd Lewis 2006). Segrss ecosystems re unique ecuse they provide structure to rren sediment ottoms, enhnce community diversity through incresed primry productivity, provide sustrte for epiphytes tht fuel intricte food wes, serve s criticl nursery hitt, nd stilize costl sediments (Beck et l. 2001, Burfeind nd Stunz 2006, Duffy 2006). Segrsses ct s importnt energetic links etween terrestril nd mrine ecosystems y exporting out qurter of their net production to djcent ecosystems (Durte nd Cerián 1996, Durte 2002). Unfortuntely mny vlule segrss ecosystems re decresing t n lrming rte worldwide from direct nd indirect humn impcts (Costnz et l. 1997, Durte 2002, Bostrom et l. 2006). Direct impcts, such s mechnicl dmge from dredging, costl construction, silttion, trwling, ot scrring nd nchoring, om lsts, eutrophiction, fish frming, nd food we ltertions, re ll visile cuses for segrss declines (Durte 2002, Bostrom et l. 2006, Erftemeijer nd Lewis 2006). Indirect cuses, such s glol climte chnge, nd nturl dissters, such s hurricnes, re lso contriuting to segrss declines (Durte 2002, Bostrom et l. 2006, Erftemeijer nd Lewis 2006). In Florid, impcts from propeller scrring re incresing nd re thretening the existence of segrss ecosystems (Hndley et l. 2007). The only sttewide ssessment of segrss scrring documented t lest 70,000 h of segrss were moderte to severely-scrred (Srgent et l. 1995). There re mny high trffic oting res round the stte of Florid tht hve een studied, including Tmp By nd the Florid Keys (Hndley et l. 2007). A propeller scr is creted when ot propeller ters through the rhizoml mt of segrss ed, llowing for erosion tht my 1

14 led to the cretion of rren snd trenches, nd thus segrss ed frgmenttion (Durko et l. 1992, Burfeind nd Stunz 2006). Segrss recovery from propeller scr dmge is oth site specific nd depth dependnt. Full recovery of shol grss, Hlodule wrightii, tkes yers (Tle 1). For turtle grss, Thlssi testudinum, complete recovery from propeller scr cn tke yers (Tle 1). Differences in recovery etween sites my e due to strength of currents, sediment type, nd wter depth (Durko et l. 1992, Dwes et l. 1997, Hmmerstrom et l. 2007). Both Tmp By (Dwes et l. 1997) nd the FL Keys (Kirsh et l. 2005) hve tidl currents tht increse erosion fter propeller scr injuries. Tidl currents cn cuse the scrs to ecome deeper trenches where segrss recoloniztion is less likely (Kirsch et l. 2005). Segrsses typiclly cnnot grow down into trenches, mking scr recovery very difficult or impossile without outside intervention (restortion) (Kirsch et l. 2005). Though propeller scr dmge hs een studied on Florid s west cost nd Keys, propeller scrring hs never een studied on the est cost. The Indin River Lgoon system (IRL) long the est cost of centrl Florid ws designted s n Estury of Ntionl Significnce y the Environmentl Protection Agency nd hs een listed s the most productive nd species-rich estury in North Americ (IRL NEP 2008) (Figure 1). Biologicl diversity in the IRL is very high in prt ecuse it is locted in trnsition zone tht encompsses temperte nd sutropicl climte zones (Stewrd et l. 2006). This diverse ssemlge provides criticl hitt for over 2000 species of mcrophytes, invertertes, fishes, irds nd mmmls (Smithsonin Institution 2007). It lso contins the most diverse ssemlge of segrss species of ny estury in the United Sttes, consisting of seven species: Hlodule wrightii, Syringodium filiforme, Thlssi testudinum, Hlophil johnsonii, Hlophil decipiens, Hlophil engelmnnii, nd Ruppi mritim (Stewrd et l. 2006). The vlue of IRL segrss eds re estimted t $329 million per yer, or $14,000 per 2

15 hectre for the commercil nd recretionl fishing industry lone (IRL NEP 2008). The vlue for segrsses is even higher if other types of recretion, esthetics, nd wter qulity functions re considered; however, there re currently no monetry estimtes for these enefits (IRL NEP 2008). Over the lst 50 yers, the undnce of segrss the IRL hs decresed 13% on verge, nd up to 90% in some res (Virnstein et l. 2007). Between 1992 nd 1993, it ws documented tht 2,400 hectres of segrss in the IRL were scrred y ots when surveyed (Srgent et l. 1995). Over the pst decde, n increse in oting ctivity led to incresed segrss scrring in the IRL (MINWR CCP 2008). Aeril surveys documented n increse in the numer oters in the northern IRL from 2002 to 2006, which ccompnied the trend of incresing humn popultion in surrounding res tht hve negtively impcted the lgoon (Scheidt nd Grreu 2007). Aout 46,000 oters were documented in the northern IRL etween 2006 nd 2007, of which 76% were fishing ots (Scheidt nd Grreu 2007). Boters used the northern IRL yer round, nd there ws no difference in the men numer of ots per month (Scheidt nd Grreu 2007). Over hlf of the oters using the IRL trveled miles to fish, showing tht the IRL is very populr plce to go oting for ll people in centrl Florid nd not just costl residents (Scheidt nd Grreu 2007). In ddition to the increse in oting ctivity, there hve een dvnces in oting technology tht llow recretionl oters to trvel with outord motors in shllower wters. Newly designed, shllow ot hulls ccompnied y outord motors mounted on hydrulic lifts re now le to mneuver in very shllow res. For exmple, Flts Cts TM sttes their shllow wter ots cn trvel nd mneuver in s little s 9 cm of wter on plne t speeds up to 32 km/h (fltscts.com/performnce.htm). 3

16 Moderte to severely-scrred res hve incresed in the northernmost portion of the IRL, lso known s Mosquito Lgoon (MINWR CCP 2008). In response to this increse, Merritt Islnd Ntionl Wildlife Refuge (MINWR) encted pole-troll zone to protect 1,272 hectres of Mosquito Lgoon segrss in the spring of 2006 (MINWR CCP 2008). The Pole-Troll Zone prohiits oters from using their min comustion engines, with the exception of mrked chnnels. Boters re llowed to troll with shllow electric motorized propeller, pole, or pddle to mneuver in plces outside the chnnels. Fines for the destruction of segrsses rnge from $ (FWC 2008). Between 2002 nd 2006, eril surveys documented n increse in oters from 483 to 603 in the Pole-Troll Zone, of which 86% were recretionl fishing ots (Scheidt nd Grreu 2007). In 2006, Scheidt nd Grreu documented tht the Pole-Troll Zone chnged oter ehvior to incresed use of poles nd trolling motors, however 80% of the oters in trnsit did not use the chnnels (2007). MINWR immeditely initilized lrge-scle nlysis of eril photogrphy (tken in the summer of 2007) to ssess the totl mount of segrss re currently dmged, the effectiveness of the Pole-Troll Zone, nd to monitor the scr recovery from repeted eril photos. This project is nering completion y Dynmc Corportion (D. Scheidt personl communiction). My study complements this reserch s ground-truthing, fine-scle comintion of oservtions nd mnipultive experiments to document: (1) the existence of multiple scr types in Mosquito Lgoon, (2) recovery time of ech type of scr, nd (3) restortion methods for the most severe scr type. My results re importnt to help develop segrss mngement pln Mosquito Lgoon. My study is novel ecuse it is the first to descrie nd document the severity of individul propeller scrs in very shllow wters of Mosquito Lgoon. Initil mesurements of existing propeller scrs were used to design mnipultive field experiment to document the nturl 4

17 recovery time of the dominnt segrss, H. wrightii in different scr severities. A second mnipultive field experiment ws conducted to test if three restortion methods could speed up the recovery time of H. wrightii when pplied to the most severe scr type. 5

18 CHAPTER 2: BASELINE SCAR MEASUREMENTS Methods Propeller scrs hd never een mesured in Mosquito Lgoon, so I gthered seline mesurements on existing ot propeller scrs in Mosquito Lgoon. My ojectives were to document differences in severity within individul scrs, nd find the men dimensions for ech scr severity. In ddition, the oservtions nd mesurements detiled elow were necessry informtion to design mnipultive experiments to exmine H. wrightii segrss recovery with nd without restortion descried in Chpter 3. Study orgnism Hlodule wrightii, is the dominnt segrss in Mosquito Lgoon (Hll et l. 2001). Also known s shol grss, H. wrightii (Figure 2), is considered pioneer species of segrss ecuse it cn tolerte greter vritions in slinity, wter depth, nd clrity thn other segrss species (Srgent et l. 1995, Dunton 1996). Compred to Thlssi testudinum, nd Ruppi mritim, H. wrightii, is the most slt tolernt nd le to survive up to 70 ppt (Koch et l. 2007). In Mosquito Lgoon, H. wrightii occur t men depth rnge of m (Hll et l. 2001). Hlodule wrightii requires pproximtely 20% direct light to grow (Stewrd et l. 2005). Hlodule wrightii cn e found on sustrte rnging from snd to mud, usully contining less thn 6% orgnic composition (Terrdos nd Durte 1999, Hemming nd Durte 2000). Hlodule wrightii nd ll segrsses re sumerged ngiosperms, underwter plnts tht spred using horizontl underground growing stem structure clled rhizome (Mrá nd Durte 1998, Hemming nd Durte 2000, Mrá et l. 2006). From the rhizome, oth roots nd leves re produced (Durte 2002). Roots re shllow nd crete n oxidized micro-lyer in which 6

19 nitrogen-fixing cteril communities reside (Welsh 2000). Leves, lso clled ldes, re undled together in vrious numers depending upon the seson, nd re ound together t the shoot. These verticl stems, or shoots, re the most commonly counted structure for segrss density mesurements. Hlodule wrightii is perennil monocot tht produces seeds tht cn lst up to 46 months (McMillin 1991, Orth et l. 2000). However, seeds or seedlings of H. wrightii hve never een documented in the IRL, nd therefore sexul reproduction is considered sent or rre event (Hll et l. 2006). The primry dispersl method for H. wrightii in the IRL is rhizome frgmenttion (Hll et l. 2006). Frgments of H. wrightii contining 3 short shoots were most commonly found floting ove segrss eds, nd were discovered to e vile for up to 4 weeks of floting in the IRL (Hll et l. 2006). In the Crien nd Florid, H. wrightii is one of the first species to colonize, followed y Syringodium filiforme, nd eventully leding to the climx species Thlssi testudinum (Mrá nd Durte 1998). These three species cn coexist in the sme segrss ed due to root prtitioning (Duffy 2006). Hlodule wrightii hs the shortest roots nd occupies the surfce sediment, S. filiforme occupies the re elow H. wrightii roots nd T. testudinum hs the deepest roots tht reside elow the other two species (Duffy 2006). Study re Mosquito Lgoon is the northern portion of the IRL long the est cost of centrl Florid (Dys 2002) (Figure 1). It stretches from Ponce de Leon Inlet south to Cpe Cnverl, Florid nd 16,000 h re protected y MINWR nd Cnverl Ntionl Seshore (CANA) (Scheidt nd Grreu 2007). Mosquito Lgoon is home to 41 federlly listed species, which is more thn ny other refuge or prk in the continentl United Sttes (MINWR CCP 2008). Mosquito Lgoon s 7

20 4,500 hectres of segrss eds re world renown for sport-fishing, nd provide hitt for commercilly importnt fish species including snook, trpon, red nd lck drum, spotted se trout, nd striped mullet (MINWR CCP 2008). Mosquito Lgoon is unique to previously studied propeller scrred res ecuse of the shllow wter depth, slow wter motion nd fine sediments (Stewrd et l. 2006). Mosquito Lgoon is very shllow, 1.7 m depth on verge (Stewrd et l. 2006). Southern Mosquito Lgoon hs semi-diurnl tides tht re microtidl, with the mximum wter level chnge of 10 cm nd less thn 2 cm in some res (Smith 1993, Hll et l. 2001). Wter motion in the lgoon is predomintely wind-driven, with winds rnging from 0-30 mph on verge tht crete wter vritions of only ±10-30 cm (Smith 1993, Hll et l. 2001). Wter levels in the lgoon re sesonl, occurring with nnul rise nd fll of ocen wter levels, with high wter levels peking in Octoer-Novemer, nd lowest wter levels in April-My (Hll et l. 2001). Averge rinfll for the lgoon is 1 cm (Hll et l. 2001). During the yer, men wter slinity rnged from ppt, men wter temperture rnged from o C, nd light ttenution rnged from m -1 with 0.92 m -1 on verge (Hll et l 2001). Sediments in Mosquito Lgoon primrily consist of fine slit-cly lom (Stewrd et l. 2006). Mesurements of existing propeller scrs To understnd the diversity of propeller scrs currently in Mosquito Lgoon, I flew in helicopter with MINWR to document res tht were severely scrred. Then I mesured ll scrs (110 totl) tht could e found y ot in ech severely scrred re long the length of Mosquito Lgoon etween Septemer 2006 nd Ferury 2007 (Figure 3). Mesurements for ech scr included: depth, width, H. wrightii cnopy height in the scr, nd H. wrightii cnopy height surrounding the scr. Other descriptive mesurements collected included: GPS loction, scr 8

21 orienttion (compss direction), slinity, temperture, wter clrity depth (mesured efore entering wter using turidity tue in cm), nd sediment from one core (3 cm dimeter x y 5 cm deep) per scr. The percent cover of the following ws lso recorded: H. wrightii cover in ech scr, H. wrightii cover surrounding ech scr, rhizome cover in ech scr, drift lge cover in ech scr, nd lef litter cover in ech scr. Percent cover ws mesured using 1 x 0.25 m strip qudrt divided into four sections, only one 0.25 x 0.25 m section in the qudrt ws counted for density mesurements, nd the qudrt ws hphzrdly chosen per scr (Dwes et l. 1997). Intr-scr mesurements Intr-scr vrition is the chnge in scr severity type within single scr. The distnce etween nd length of ech scr type (severity) ws mesured for 15 m in one direction for ll 110 scrs. For every severity scr type encountered within one scr, the following mesurements were collected: scr depth, width, nd cnopy height in nd surrounding the scr, nd percent cover of H. wrightii in scr, H. wrightii surrounding scr, rhizome in scr, nd the numer H. wrightii shoots in nd surrounding the scr. Sediment nlyses Sediment cores collected from ech scr were nlyzed for grin size frctions nd orgnic content. All 110 sediment smples were dried for 48 hours t 80 o C nd then ground with mortr nd pestle. Smples were then sieved into 5 frctions: > 5 mm, 5 mm, 2 mm, 1 mm, nd < 1 mm. Ech frction ws weighed nd percents of ech frction were clculted. Next, sediment grin size frctions were recomined, nd the complete sediment smples were dried t 80 o C for 24 hours prior to eing plced in muffle furnce for orgnic content nlysis (Prker 1983, Fino et l. 1995). Smples were weighed efore nd fter eing plced 9

22 in muffle furnce for 2 hours t 450 o C (Prker 1983, Fino et l. 1995). The orgnic content ws clculted from the chnge in weights nd expressed s percent orgnic mtter. Sttisticl nlyses All mesurement dt were initilly tested for normlity using Shpiro-Wilks test, nd homogeneity of vrince using Levene s test in SPSS Sttisticl Softwre (Version 11.5). All mesurements were trnsformed using log 10, except for percent frctions which were trnsformed using rc sin to normlize the dt. Only trnsformed mesurement dt were used in nlyses. To document if there were different types of scrs, I entered ll mesured vriles in Cnonicl Discriminnt Anlysis to determine the numer of unique scr types nd which vriles were most importnt in clssifying ech scr type. The Wilk s Lmd test ws used to test which mesurements were significnt in defining scr groups. The frequency of ech scr type ws clculted from the 110 mesured scrs. The overll men nd men per scr type for ech generl mesurement were clculted. Correltions of ll mesured vriles were tested with Person two-tiled tests. To further test the significnt correltion of scr depth versus scr width, liner regression ws clculted for scr Type 1, since it ws the only scr type hving depth. All scrs were mpped in GIS, nd GPS coordintes were used to compute distnces in meters from ech scr to the nerest chnnel (Figures 1, 3). Chnnels included the Atlntic Intrcostl Wterwy, the edge of the Pole-Troll Zone nd the inner chnnels of the Pole-Troll Zone. Distnces were used to see if there ws pttern in scr sediment type using Cnonicl Discriminnt Anlysis. Sediment frctions were lso tested to see if there ws pttern mong scr types nd if sediment type could predict scr types using Cnonicl Discriminnt Anlysis. 10

23 Sediment frctions for ech scr type were tested with one wy nlysis of vrince (ANOVA) to test if the men of ech frction ws significntly different in ech scr type. Results Scr severity ctegories Cnonicl Discriminnt Anlysis ws used to test for ctegories of scr severity sed on mesured scr dimensions, nd four groups were clerly distinguished (Figure 4). Types 1 nd 4 scrs (most extreme) were more ccurtely clssified nd intermedite scr Types 2 nd 3 were less cler. (Tle 2). I concluded tht four types of propeller scrs ccurtely represented propeller scrring in Mosquito Lgoon ecuse 91.8% of ll scrs were correctly clssified (Tle 2). Significnt vriles used to clssify scr types were: (1) shoot density in scr, (2) percent root cover, (3) scr depth, nd (4) scr cnopy height (Wilk s Lmd p < 0.001). Type 1 ws the most severe scr type, exhiiting trench formtion, hving the ottom of the scr deeper thn the surrounding sediment surfce nd n sence of segrss shoots nd visile rhizomes within the scr (Figure 5). Type 2 ws scr type hving the ottom level with the surrounding sediment surfce, nd with n sence of segrss shoots nd visile rhizomes (Figure 5). Type 3 ws scr type hving the ottom level with the surrounding sediment surfce, nd contining rhizomes ut no shoots (Figure 5). Type 4 ws scr hving ottom level with the surrounding sediment surfce, nd contining rhizomes with shoots tht were cut shorter thn the surrounding cnopy height (Figure 5). All four scr types exhiited liner pth through the segrss nd could not e distinguished from ot. However, eril photos were underestimting the mount of scrring ecuse Type 4 scrs could not e seen, since there ws not enough contrst etween the segrss nd the sediment (D. Scheidt personl communiction). Out of the 110 scrs 11

24 mesured, the most severe scr type (Type 1) ws the most common t 56.4%, followed y the lest severe scr type (Type 4) t 27.3%, Type 2 t 10.0%, nd Type 3 t 6.4% (Figure 6). Generl scr mesurements The men width (± S.E.) of ll mesured scr ws 32 (± 2) cm, nd the men depth of scrs rnged from 9 (± 1) cm in Type 1 to 1 (± 1) cm in Type 4 (Tle 3). Scr depth ws significntly correlted with scr width in Type 1 scr (ANOVA: p < 0.001, r 2 = 0.476; Figure 7). Scr direction ws not significntly correlted with ny mesured vrile (Person correltion: p > 0.088). Two species of segrss, Hlodule wrightii nd Syringodium filiforme, were found in nd surrounding the scrs. The men shoot density of S. filiforme ws 1%, while H. wrightii comprised 99% of ll mesured segrss. Since H. wrightii ws the dominnt species of segrss in Mosquito Lgoon, from here on I report only segrss mesurements for H. wrightii. The men H. wrightii shoot count in scrs rnged from 0 (± 0) in Type 1 to 44 (± 6) in Type 4 scrs. Scr Type 4 men H. wrightii shoot density inside the scr ws less thn hlf of the men surrounding H. wrightii shoot density (Tle 4). The men H. wrightii cnopy height in scrs rnged from 0 (± 0) in Type 1 scrs to 12 (± 1) in Type 4 scrs. The surrounding segrss cnopy height men ws similr in ech scr type, rnging from cm (Tle 3). The percent cover (± S.E.) of H. wrightii in Type 4 scrs ws 25 (± 4)% nd ws less thn hlf of the men percent coverge of H. wrightii surrounding the scr, 62 (± 5)% (Tle 4). Men percent root cover rnged from 56-73% in scr Types 3 nd 4, while scr Types 1 nd 2 did not hve roots. Men percent lef litter cover ws quite vrile, rnging from 10 (± 4)% in scr Type 4 to 55 (± 5)% in scr Type 1 (Tle 4). Men percent cover of drift lge ws lso highly vrile mong scr types (Tle 4). 12

25 Intr-scr mesurements Of the 110 scrs mesured, only 2 scrs contined multiple scr types. Both scrs lternted etween Type 1 nd Type 4 scr severities, repetedly lternting long the 15 m sections mesured. Scr Type 1 ws more frequent comprising 65% of oth scrs, while Type 4 severity comprised 35%. Severities switched 5 times in the first scr nd 6 times in the second scr. The men length (± S.E.) of severity section ws 1.6 (± 0.4) m. The men length rnged from 1.9 (± 0.6) m in Type 1 scr sections to 1.3 (± 0.6) m in Type 4 sections. Only Type 1 scrs sections hd depth, nd the men depth (± S.E.) ws 13.9 (± 1.7) cm. The remining 108 scrs hd constnt severity for t lest 15 m long the length of ech scr. Scr sediment nlysis Sediment composition vried mong scr types. Scr Types 1 nd 4 were significntly different in ll frctions (ANOVA: p < 0.01; Figure 8). Type 4 scrs were dominted more y fine (< 1 mm) sediments thn Type 1 scrs, with intermedite levels in scr Types 2 nd 3. Overll, 60% of ll sediment smples were predomintely composed of the most fine sediment (< 1 mm) frction (Figure 8). Men orgnic content ws 2.3 (± 0.1)% for ll smples nd did not significntly differ mong scr types. Cnonicl Discriminnt Anlysis ws used to test if scr ctegories could e clssified with sediment type nd loction, using the following mesurements: sediment frctions, nerest chnnel distnce, scr GPS loction, scr depth, nd wter depth. There ws no pttern in the loction of sediment types (Figure 9). Sediment frctions clssified scrs into two groups rther thn four groups, s predicted y the scr nd segrss mesurements (Tle 5, Figure 9). Thus, sediment frctions were not good predictors of the scr types. 13

26 Discussion My study ws the first to discover tht different severities within individul propeller scrs exist, nd the first to document the dimensions of propeller scrs in the IRL. Of the scr types I identified, only Types 1 nd 2 were documented in previous studies (Tle 1). The scrring oserved in Mosquito Lgoon is less severe thn other res, such s Tmp By nd the Florid Keys. Segrss recovery from propeller scr dmge is oth site specific nd depth dependnt (Tle 1). Sites tht hve more tidl or current motion re likely to erode scrs deeper nd wider, mking segrss recovery difficult (Kirsh et l. 2005). Sediment composition my influence how well the segrss cn grow nd remin rooted when exposed to currents (Durko et l. 1992, Dwes et l. 1997, Hmmerstrom et l. 2007). For exmple, Tmp By s sustrte is composed of siliceous snd tht is firmer thn the Florid Keys, which hs different minerl composition in the sustrte of cronte snd (Fonsec et l. 2004). Scrs in the Florid Keys hve more corse sediments with lower ph thn the sediments from surrounding segrss eds; this however ws not found in Tmp By scrs (Srgent et l. 1995). Sites tht re in more shllow wter my tke longer to recover (Tle 1). Mosquito Lgoon is microtidl with very slow, wind-driven currents. I predict scrs creted in Mosquito Lgoon to e less severe thn previously studied res. However, Mosquito Lgoon is lso very shllow with fine, unstle sediments, which my mke the recovery more difficult. Tle 1 shows propeller scrs cn rnge m in width, nd cm in depth. Previously documented scrs cn e fr more severe thn the most severe scrs (Type 1) I hve documented in Mosquito Lgoon, with men size of 32 cm wide x 9 cm deep. The mesurements of existing scrs were nlyzed to not only understnd the severity of propeller scrring in Mosquito Lgoon, ut to design experiments to test the recovery of 14

27 H. wrightii from ll propeller scr types. The first ojective in plnning the experiments ws to document if different scr types exist, nd the frequency of ech type in Mosquito Lgoon. Four different scr severities were documented (Figure 5). Out of the 110 scrs mesured in the field, the most severe scr type ws the most common (Figure 6). My second ojective ws to identify men mesurements for ech scr type, nd use them to mke experimentl scrs. Only the men dimensions from the four scr types were used in testing the recovery time of H. wrightii. Multi-severity scrs were not included in the experiments ecuse they were very rre in the lndscpe, only 2 scrs out of 110 were found to hve multiple severities. Since 99% of ll segrss mesured ws H. wrightii, it ws the only segrss mesured for recovery time in the experiments. My third ojective ws to determine if sediment composition ws n importnt fctor in identifying scr types, nd if sediment types were distriuted in pttern or grdient in the lndscpe. The distriution of sediment types ws used to decide if the experiment should e plnned in more thn one loction. Cnonicl Discriminnt Anlysis results showed tht sediment type ws not good predictor of scr type, nd no cler sediment ptterns were found in the lndscpe. Sediments were heterogeneously distriuted, nd were predominntly composed of fine sediments (< 1 mm), with 2% men orgnic mtter (Figure 7). Since sediment types were not found in specific pttern or t specific loctions, only one loction ws used for the mnipultive experiments. In summry, my study documented propeller scr dimensions nd individul scr severities for the first time in Mosquito Lgoon. I found tht there were four min res with intense scrring, three of which were within the Pole-Troll Zone oundries (Figure 3). All scrs in these res were mesured, nd the most common scrs found were the most severe scrs (Type 1). If 15

28 regultions were strictly enforced, then the mjority of the scrs in the segrss would e mde y trolling motors. Trolling motors hve djustle mounts for chnging propeller depth in shllow wter to void hitting the sediment. Trolling motors re ttery powered, nd re not powerful enough to plow through the sediment creting less severe scrs (Type 4). According to Engemn et l. (2008), segrss restortions re fr more expensive thn prevention methods tht involve dding signge nd lw enforcement officers. I would recommend tht the Pole-Troll Zone should hve more thn 2 signs ( sign t ech chnnel entrnce), especilly since Scheidt nd Grreu found tht 80% of oters in trnsit do not use the mrked chnnels (2007). New signs would only cost n verge of $192 ech (Engemn et l. 2008). I lso recommend tht there e more thn one officer to ptrol the 1200 h Pole-Troll Zone. More ptrol officers would ech cost n verge of $59,400 per yer (Engemn et l. 2008). To put costs into perspective, single ptrol officer s nnul slry ws equivlent to the vlue of 0.42 h of segrss in the Florid Keys (Engemn et l. 2008). For every segrss hectre in the IRL protected to llow recovery, or protected to prevention ot scrring, n dditionl $14,000 is gined in fisheries resources (IRL NEP 2008). More thn 119 species of fishes hve een documented in the IRL, nd segrss percent cover is n importnt fctor in determining fish ssemlges (Kupschus nd Tremin 2001). Mny fish species in the lgoon reproduce in the ocen, ut rely on esturine juvenile phse to complete the reproductive cycle (Kupschus nd Tremin 2001). Segrss ecosystems should e protected ecuse they re criticl in creting hitt in res of rren sediment, enhncing community diversity, estlishing intricte food wes, nd stilizing costl sediments (Beck et l. 2001, Burfeind nd Stunz 2006, Duffy 2006). Segrss eds in the Indin River Lgoon provide criticl hitt for over 2000 species of invertertes, fishes, irds nd mmmls, nd re key 16

29 reson why the IRL is the most productive nd species-rich esturies in North Americ (IRL NEP 2008, Smithsonin Institution 2007). 17

30 CHAPTER 3: MANIPULATIVE EXPERIMENTS Methods Using the men mesurements for ech scr type from Chpter 2, two complimentry mnipultive experiments were conducted to exmine nd potentilly mximize Hlodule wrightii recovery from propeller scrs in shllow Mosquito Lgoon wters. The first experiment evluted the nturl recovery rte of H. wrightii in the four scr types. I hypothesized tht different severity in individul propeller scrs would cuse different recovery times for H. wrightii. The second experiment ssessed the success of known restortion methods for the most severe scr type (Type 1). I hypothesized tht the ddition of different restortion methods would cuse Type 1 scrs to recover fster thn nturl recovery. Study re The loction for my mnipultive experiments ws in n re with miniml ot trffic. In order to reduce the likelihood of ots running over the experiments, experiments were plced in the Pole-Troll Zone where oters were not llowed to use their min comustion engines (Figure 1). Experiments were plced in dense segrss ed etween the rrier islnd of Cnverl Ntionl Seshore nd n islnd west of the rrier islnd within the Pole-Troll Zone, which effectively locked ot trffic in two directions (Figure 1). The segrss ed s men depth (± S.E.) ws 56 ± 4 cm. The segrss ed ws dominted y H. wrightii, with occsionl smll ptches of Ruppi mritim in spring (personl oservtion). The GPS coordintes for this segrss ed loction were N nd W

31 Lyout nd monitoring of experiments The experimentl tretments descried elow were plced in rndomized lock design to control for nturl lndscpe vrition (Figure 10). For the rndomized lock design, one replicte of ech tretment ws plced in lock, with totl of eight tretments per lock (Figures 10, 11). There were 10 replictes for ech tretment, nd thus 10 locks totl. Ech lock ws 18 x 18 m nd ws divided into nine equl 6 x 6 m sections. Eight of the nine sections contined single scr tretment (Figure 10). In the center of ech section, ech experimentl scr ws plced minimum of 2 m wy from ll other scrs. All experimentl scrs were 200 cm long x 32 cm wide. Tretment loction nd scr direction within ech lock were rndomized using Excel. Ech experimentl scr ws mrked with n luminum tg on PVC pole one meter wy from ech end. All PVC poles were pushed down into the sediment leving 15 cm remining ove the sediment surfce (Figure 12). All experimentl scrs were mesured in the center for new segrss recruitment nd growth. Mesurements included scr width, depth, scr slope ngle, H. wrightii cnopy height inside scr, H. wrightii cnopy 1 m wy from scr side, H. wrightii shoot density per 25 x 25 cm inside scr, nd H. wrightii shoot density per 25 x 25 cm t 1 m wy from scr side. Also, the percent cover of H. wrightii inside the scr, the percent cover of H. wrightii 1 m wy from the scr, percent lef litter cover, nd drift lge cover were mesured. Wter mesurements were collected on ech smpling dte nd included: slinity, temperture, nd wter clrity depth (mesured efore entering wter with turidity tue in cm). Also, ny erosion tht occurred in the scrs ws documented. All mesurements were tken immeditely prior to digging the experimentl scrs, then fter 2 weeks, 1 month, 2 months, 3 months, 6 months, 9 months, nd 12 19

32 months. The initil mesurements were tken on 17 July 2007, nd finl mesurements were tken on 3 August Nturl recovery of segrss in experimentl propeller scrs To document the nturl recovery time of H. wrightii in ech scr type, 10 replictes of ech type of scr were creted nd mesured for recoloniztion t set intervls over one yer. Tretments included: (1) Type 1 scrs with depth of 9 cm, no roots nd no shoots inside the scr, (2) Type 2 scrs with no depth, no roots, nd no shoots, (3) Type 3 scrs with no depth nd no shoots, ut with intct roots, nd (4) Type 4 with roots, shoots, nd ldes cut from men cnopy height of 44 cm to 12 cm high. The controls for this experiment were res of dense segrss of the sme dimensions (200 cm long x 32 cm wide) s the experimentl scrs within lrger segrss eds. Type 1 scrs were creted using 32 cm PVC plow, so tht the depth (9 cm), width (32 cm), nd scr ngle (45 o ) in ll experimentl Type 1 scrs remined constnt (Tle 6). The plow consisted of digonlly cut nd shrpened piece of 32 cm dimeter PVC with luminum mesh ttched t the end with rivets, nd fitted with n luminum rod through the center for hndles (Figure 13). The PVC plow creted uniform scr depth nd slope ngle for ll Type 1 scrs, which could not e done using hnd shovel or ot propeller (Tle 6). Type 2 scrs were creted y severing the segrss roots with grden spde nd rking out the segrss inside scrs with 32 cm long rke. Type 3 scrs were creted y trimming ll ove-ground iomss to the enthos within 200 cm x 32 cm res using grss shers. Type 4 scrs were creted y trimming ll segrss ldes to 12 cm ove the enthos within 200 cm x 32 cm res using grss shers (Tle 6). All mesurement dt were initilly tested for normlity using Shpiro-Wilks test, nd homogeneity of vrince using Levene s test. Due to non-normlity, experiments were nlyzed 20

33 using seprte Kruskl-Wllis tests (the non-prmetric equivlent of one-wy ANOVA) to test for significnt differences mong tretment mens for ech growth mesurement nd scr dimension vrile seprtely for ech time period. Since Kruskl-Wllis tests were repeted eight times for ech mesured vrile, the Bonferroni method ws used to correct the significnce level from p= 0.05 to p = Restortion of experimentl propeller scrs To test restortion methods to potentilly reduce the recovery time of H. wrightii, I tested three known protocols with the most severe scr type nd mesured recoloniztion t set intervls over one yer. Hlodule wrightii is considered the idel trnsplnt species since it is more tolernt to wide rnge of conditions thn other segrss species (Buckholder et l. 1994; Dunton 1996). Hlodule wrightii cn recover fster thn other species of segrss ecuse it hs greter density of shoot nd rhizome nodes, providing the ility to rnch more often nd produce more shoots with ldes (Durko et l. 1992; Srgent et l. 1995). Hlodule wrightii is often used to stilize the sediment for climx segrss species such s T. testutinum; this technique is clled compressed succession (Hll et l. 2006). Hlodule wrightii hs een successfully replnted using stples (Hll et l. 2006), pet pots (Sheridn et l. 1998), plnted in snd tues (Hll et l. 2006) nd lrge sod squres (Thorhug 1986). I decided to use restortion methods shown to e successful trnsplnting H. wrightii in other res of Florid. Hll et l. ws successful in using the stple method to trnsplnt H. wrightii nd snd tues to fill in scrs in oth Tmp By nd the Florid Keys (2006). The scrs in oth plces recovered in one yer to control H. wrightii densities (Hll et l. 2006). Since this ws successful, I decided to use the stple method to trnsplnt H. wrightii, nd use fine snd to fill scrs. I did not use snd tues 21

34 ecuse commercilly ville tues were lrger thn the experimentl scrs were deep, nd the current flow ws miniml in Mosquito Lgoon. All tretments nd the control strted y creting Type 1 scr trench (200 x 32 x 9 cm). Restortion methods tested were: (1) plnting H. wrightii in the scr trench, (2) filling the scr trench with commercilly ville snd, nd (3) filling the scr trench with commercilly ville snd followed y plnting H. wrightii. The control ws Type 1 scr with no dded restortion. All tretments nd the control were replicted 10 times. For the fill only nd the fill nd plnt tretments, scrs were filled with 0.57 m 3 of Kolorscpe TM extr fine ply snd (1 mm grin size). Ech g ws opened underwter nd poured directly into the scrs. After ll gs were poured, the scr surfces were smoothed flt y hnd. All plnting of H.wrightii used the stple method descried in NOAA s Guidelines for the Conservtion nd Restortion of Segrsses in the United Sttes nd Adjcent Wters (Fonsec et l. 1998). Ech trnsplnt unit of H. wrightii ws composed of three-shoot frgment with growing tip. The rhizome of ech trnsplnt ws ttched to metl grden stple with pper twist tie (Figure 14). Frgments with three shoots were used ecuse it is the most common frgment size produced from nturl H. wrightii frgmenttion in the IRL (Hll et l. 2006). Three-shoot frgments cn ttch to sustrte nd root within 2 weeks (Hll et l. 2006). Ech plnted scr contined 16 trnsplnt units, in 4 x 4 rry. Ech row ws spced pproximtely 25 cm prt within the scr. All segrss ws collected nd trnsplnted within 4 hours. All mesurement dt were initilly tested for normlity using Shpiro-Wilks tests, nd homogeneity of vrince using Levene s tests. Due to non-normlity, experiments were nlyzed using seprte Kruskl-Wllis tests to test for significnt differences mong tretment mens for ech growth mesurement nd scr dimension vrile seprtely for ech time period. Since 22

35 Kruskl-Wllis tests were repeted eight times for ech mesured vrile, the Bonferroni method ws used to correct the significnce level from 0.05 to Sediment core nlysis One sediment smple per experimentl scr ws collected efore the cretion of the scr nd t the end of the experiments to test if composition chnged over time. Ech sediment core ws nlyzed for grin size frctions nd orgnic content. All sediment smples were dried for 48 hours t 80 o C nd then ground with mortr nd pestle. Smples were sieved into five frctions: > 5 mm, 5 mm, 2 mm, 1 mm, nd < 1 mm. Ech frction ws weighed nd percentges were clculted. After sieving, complete sediment smples were dried t 80 o C for 24 hours prior to eing plced in muffle furnce for orgnic content nlysis (Prker 1983, Fino et l. 1995). Smples were weighed efore nd fter eing plced in muffle furnce for 2 hours t 450 o C (Prker 1983, Fino et l. 1995). The orgnic content ws clculted from the chnge in weights nd expressed s percent orgnic mtter. Kruskl-Wllis tests were used to document if there were significnt differences mong initil nd finl sediment frctions, nd orgnic content in smples. Results Site consistency Wter temperture rnged from o, slinity rnged from ppt, nd wter clrity depth rnged from cm mong locks (Figures 15-17). The men wter depth (± S.E.) ws 56 (± 4) cm for ll locks (Figure 18). Wter clrity my hve een greter if mesured with different device ecuse my turidity tue ws 60 cm long; the mximum mesurement possile ws 60 cm (Figure 17). 23

36 Hlodule wrightii density, percent cover nd cnopy height of the segrss surrounding scrs were not significntly different from the etween locks over one yer (Figures 19-21). Men density (± S.E.) of H. wrightii surrounding scrs nd in control plots rnged from 77 (± 4) to 235 (± 10), however some res hd densities s high s 469 shoots per 25 x 25 cm in the spring (Tles 6, 8). Men cnopy height (± S.E.) of H. wrightii surrounding scrs nd in control plots rnged from 21 (± 1) to 47 (± 1) cm over one yer (Tles 6, 8). Men percent cover (± S.E.) of H. wrightii surrounding scrs nd in control plots rnged from 28 (± 2) to 87 (± 2)%, including hlf of the scrs in the spring nd summer hving 100% cover (Tles 6, 8). Nturl recovery experiment Type 4 scrs recovered to the H. wrightii density, percent cover, nd cnopy height of the control tretment in 3 months, nd ws the only scr type to completely recover within one yer (Figures 22-24, Tle 6). First coloniztion y single rhizome cross the width of the most severe (Type 1) scrs ws oserved t nine months. After 12 months, men H. wrightii shoot density (± S.E.) incresed from zero to 116 (± 21) in Type 1, 115 (± 20) in Type 2, nd 154 (± 31) in Type 3 scrs (Figure 22, Tle 6). Types 1, 2, nd 3 were significntly different from the control t 12 months (Kruskl-Wllis: p < 0.001; Tle 7). If scrs continue to increse in shoot density t the sme rte, then I would expect to see recovery to control density in 28 months for Type 1 nd 2 scrs, nd 21 months in Type 3 scrs. Percent cover of H. wrightii (± S.E.) incresed from zero initil cover to 46 (± 9)% in Type 1, 47 (± 9)% in Type 2, nd 58 (±12)% in Type 3 scrs t 12 months (Figure 23). Scr Types 1, 2, nd 3 were significntly different from the control t 12 months (Kruskl-Wllis: p < 0.001; Tle 7). If scrs continue to increse in percent cover t the sme rte, then I would expect to see recovery to control levels in 25 months for Type 1 nd 2 scrs, nd 20 months in Type 3 scrs. 24

37 Scr cnopy height for Types 1, 2, nd 3 ws not significntly different from the men control height (± S.E.) of 38 (± 2) cm (Kruskl-Wllis: p = 0.001; Tle 7). However, ll incresed in men cnopy height (± S.E.) from initil vlues of zero to 25 (± 2) cm in Type 1, 26 (± 3) cm in Type 2, nd 30 (± 2) cm in Type 3 (Figure 24). If scrs continue to increse in cnopy height t the sme rte, then I would expect to see recovery to control vlues in 18 months for Type 1 nd 2 scrs, nd 15 months in Type 3 scrs. There ws significnt difference etween scr root density for scr Types 1-3 nd control nd Type 4 t one yer (Tle 7). Men percent root cover (± S.E.) in Type 4 scrs recovered to 96 (± 3)%, Type 2 to 57 (± 12)%, Type 3 to 53 (± 10)%, nd Type 4 to 54 (± 11)% (Figure 25). In the recoloniztion of scrs tht hd depth, lef litter cted s nturl fill for rhizomes. Rhizomes from the surrounding H. wrightii extended towrd the center of the scr nd ttched on top of the lef litter. Lef litter percent cover ws higher in the scr Types 1-3 thn control nd Type 4 (Figure 26). Type 1 scrs initilly hd significntly greter proportion of lef litter fter the second month thn ll other tretments (Kruskl-Wllis: p < 0.001; Tle 7). At 12 months, percent lef litter cover ws low nd there ws no significnt difference in lef litter cover mong ll tretments. Scr dimension mesurements were compred with control vlues of zero depth, zero width, nd zero scr ngle. Only Type 4 scrs were not significntly different from the control in ll mesurements fter one yer (Tle 7). Although significntly different from the control fter one yer, the men scr depth ws reduced from 9 cm initilly to 4 (± 1) cm in Type 1 scrs, nd incresed to 2 (± 1) cm in Type 2 nd 3 scrs (Kruskl-Wllis: p = 0.001; Figure 27, Tle 7). If scrs continue to decrese in scr depth t the sme rte, then I would expect to see recovery in 22 months for Type 1 scrs, nd 15 months for Type 2 nd 3 scrs. Scr width lso decresed from 32 25

38 cm to 22 (± 3) cm in Type 1, 18 (± 3) cm in Type 2, nd 20 (± 7) cm in Type 3. All were significntly different from the control t the end of one yer (Kruskl-Wllis: p < 0.001; Figure 28, Tle 7). If scrs continue to decrese in scr width t the sme rte, then I would expect to see recovery in 38 months for Type 1 scrs, 27 months for Type 2 scrs, nd 32 months for Type 3 scrs. Scr slopes in Type 2-4 scrs were not significntly different from the control fter one yer. Type 1 scrs were significntly different from the control, however the slope ws reduced from 45 o to 13 (± 3) o over 1 yer (Kruskl-Wllis: p < 0.001; Figure 29, Tle 7). If scrs continue to decrese in scr ngle t the sme rte, then I would expect to see recovery in 17 months for Type 1 scrs. Sediment frctions mong scr Types 1-4 nd control were not significntly different t 1 yer. There ws significnt difference etween initil nd finl sediment in frction sizes: > 5 mm, 5 mm, 1 mm, nd < 1 mm (Kruskl-Wllis: p < 0.007; Figure 30). Sediment frctions sizes 1 mm, nd < 1 mm were the lrgest in ll smples, with ll other sizes composing less thn 10% of the smple. Initil fine sediments rnged from % in size < 1 mm, nd 23-87% in size 1 mm. Initil sediment smples men percent frction sizes (± S.E.) were 0.8 (± 0.1)% for > 5 mm, 5.7 (± 0.4)% for 5 mm, 10.3 (± 0.4)% for 2 mm, 31.7 (± 1.3)% for 1 mm, nd 51.4 (± 1.3)% for < 1 mm. Finl fine sediments rnged from 30-63% in size < 1 mm, nd 26-50% in size 1 mm. Finl sediment smples men percent frction sizes (± S.E.) were 2.0 (± 0.2)% for > 5 mm, 8.3 (± 0.5)% for 5 mm, 10.5 (± 0.3)% for 2 mm, 35.3 (± 0.8)% for 1 mm, nd 43.9 (± 1.1)% for < 1 mm. There ws no significnt difference in orgnic mong ll finl sediment smples, nd etween initil nd finl sediment smples (Figure 31). Men percent orgnic mtter (± S.E.) in initil smples ws 4.9 (± 0.4)%, nd 3.8 (± 0.1)% in finl smples. 26

39 Sesonl vritions in nturl recovery Sesonl vritions were oserved in H. wrightii density. Initilly, shoot density slightly incresed in controls nd ll scr tretments until fll Novemer 2007, then dropped to significntly lower shoot density in winter (Ferury 2008) (Kruskl-Wllis: p < 0.001; Figure 22). The lrge die-ck during winter reduced densities to 6 (± 2) shoots per 25 x 25 cm in Type 1 scrs, 13 (± 5) in Type 2 scrs, 28 (± 9) in Type 3 scrs, 73 (± 12) in Type 4 scrs, nd 80 (± 28) in the control (Figure 22, Tle 6). However, H. wrightii incresed exponentilly in ll scr types during spring nd summer (Ferury through August 2008), ending with men densities (± S.E.) of 116 (± 21) shoots per 25 x 25 cm in Type 1 scrs, 115 (± 20) in Type 2 scrs, 154 (± 31) in Type 3 scrs, 244 (± 21) in Type 4 scrs, nd 271 (± 16) in the control segrss plots (Figure 22, Tle 6). Men shoot density mong tretments in winter ws significntly lower in ll tretments thn in summer 2008 (Kruskl-Wllis: p < 0.001; Figure 22, Tle 6). A similr sesonl pttern of significntly lower men H. wrightii percent cover (Kruskl-Wllis: p < 0.001) nd cnopy height (Kruskl-Wllis: p < 0.001) ws oserved mong ll tretments. Sesonl vritions in scr dimensions were lso oserved. Scrs incresed in depth during the fll nd mintined significntly deeper depth of 12 (± 1) cm in Type 1, 7 (± 1) cm in Type 2, nd 5 (± 2) cm in Type 3 during the winter (Kruskl-Wllis: p < 0.001; Figure 27, Tle 6). Scr width erosion ws worst in winter, scrs significntly incresed in width to 41 (± 4) cm in Type 1 scrs, to 51 (± 12) cm in Type 2 scrs, 32 (± 9) cm in Type 3 scrs (Kruskl-Wllis: p < 0.001; Figure 28, Tle 6). Type 4 scrs were not eroded. The slope ngle eroded creting significnt increse scr ngle slope of 29 (± 9) o in Type 2, nd 12 (± 7) o in Type 3 over 12 months (Kruskl- Wllis: p < 0.02; Figure 29, Tle 65). 27

40 Restortion experiment There ws 100% mortlity of the 320 trnsplnts fter 60 dys (4 oservtion periods) (Tle 9). The two tretments hving H. wrightii plnting units lost lmost hlf of the plnting units fter 2 weeks, nd ll fter 2 months. Trnsplnts on stples were lost from either sediment uril or from sediment erosion under stples. Mny of the plnting units were uried in the scrs, possily due to the stple sinking through the fine sediment, or fine sediments settling out of the wter column. After 2 weeks, out hlf of the plnting units in ech plnt-only tretment were uried to men depth (± S.E.) of 6 (±2) cm. Mny trnsplnts lived for short period (up to 2 months) while ttched to n eroded or elevted stple, however ll trnsplnts died fter 2 months. Mny of the grden stples with trnsplnts in oth the plnt-only nd fill nd plnt tretments hd the sediment eroded out from under the stple exposing the trnsplnts etween 2 nd 12 cm ove the enthos. Despite none of the trnsplnts contriuting to scr coloniztion, the fill nd plnt, nd plnt-only tretments were colonized nturlly from the djcent segrss ed (Figure 32). The plnt-only tretment showed initil segrss recoloniztion fter 2 months despite the men scr depth of 10.5 cm (Tle 8). The fill nd plnt tretment did not show ny sign of recoloniztion fter 3 months (Tle 8). The snd fill tretment survived the longest, with 5 of 20 scrs retining snd throughout the entire yer (Tle 9). The snd fill tretment did provide stle sediment for surrounding H. wrightii to colonize from the djcent segrss ed. None of the restortion tretments significntly differed from the nturl recovery of Type 1 scrs for ll H. wrightii mesurements, including H. wrightii shoot density, percent H. wrightii cover, cnopy height, percent root cover, nd percent lef litter cover (Figures 32-36, Tles 8, 10). Scr depth, width, nd ngle were reduced over time in ll restortion tretments, 28

41 nd none of the tretments significntly differed from Type 1 scrs (Figures 37-39, Tles 8, 10). Only the segrss control ws significntly different from ll restortion tretments nd Type 1 scrs in ll mesured vriles (Tle 10). Sediment frctions mong restortion tretments, Type 1 scrs nd control were significntly different in frction size 5 mm fter 1 yer (Kruskl-Wllis: p = 0.031; Figure 40). There ws significnt difference etween initil nd finl sediment in frction sizes: > 5 mm, 5 mm, 1 mm, nd < 1 mm (Kruskl-Wllis: p < 0.002; Figure 40). Sediment frctions sizes 1 mm, nd < 1 mm were the lrgest in ll smples, with ll other sizes composing less thn 10% of the smple. Initil fine sediments rnged from % in size < 1 mm, nd 23-87% in size 1 mm. Initil sediment smples men percent frction sizes (± S.E.) were 0.8 (± 0.1)% for > 5 mm, 5.7 (± 0.4)% for 5 mm, 10.3 (± 0.4)% for 2 mm, 31.7 (± 1.3)% for 1 mm, nd 51.4 (± 1.3)% for < 1 mm. Finl fine sediments rnged from 13-55% in size < 1 mm, nd 25-81% in size 1 mm. Finl sediment smples men percent frction sizes (± S.E.) were 2.5 (± 0.3)% for > 5 mm, 8.1 (± 0.7)% for 5 mm, 9.7 (± 0.5)% for 2 mm, 40.5 (± 2.1)% for 1 mm, nd 39.4 (± 1.5)% for < 1 mm. There ws significnt difference in orgnic mong restortion tretments, Type 1 scrs nd control sediment smples (Kruskl-Wllis: p = 0.041; Figure 41). Also there ws significnt difference in orgnic content etween initil nd finl sediment smples (Kruskl-Wllis: p = 0.028; Figure 41). Men percent orgnic mtter (± S.E.) in initil smples ws 4.9 (± 0.4)%, nd 3.5 (± 0.2)% in finl smples. Erosion oservtions Mny of the scrs in oth experiments suffered from erosion t the scr edges, wy from the scr center where mesurements were tken. Erosion t scr edges ws seprted into three ctegories: (1) holes, (2) wsh-outs, nd (3) snd ptches (Figure 42). A hole ws defined s ny 29

42 circulr erosion tht occurred within, nd did not extend eyond the initil dimensions of the width or length of scr edge (Figure 42). A wsh-out ws type of erosion tht extended eyond the width or length of the scr nd removed H. wrightii on 1 to 3 sides surrounding the scr (Figure 42). A snd ptch ws the result of losing the top fine-sediment lyer tht creted depths rnging cm deep, long with losing ll four sides of H. wrightii surrounding the scr to erosionl forces (Figure 42). Overll, men scrs tht hd holes ws 11, men scrs tht were wshed out ws 8, nd men scrs tht were turned into snd ptches ws 2 for ech mesurement period (Tle 11). Out of oth experiments, the control nd Type 4 scrs hd significntly fewer holes nd wshouts (Kruskl-Wllis: p < 0.001). The fill-only nd fill nd plnt tretments hd the highest, hving three times more holes nd wshouts thn ll other scr tretments (Tle 11). The fill only nd fill nd plnt tretments lso hd significntly more scrs tht turned into snd ptches (Kruskl-Wllis: p = 0.007). Some scrs likely strted with one or more holes, possily cused y stingrys, which expnded into wsh out nd over time turned into snd ptch (Figure 43). Bioturtion from stingrys ws not the only erosionl force; it is likely tht dditionl ot scrring nd ot trffic wkes lso contriuted to scr erosion (Figure 44). Every time I mesured the experiments, I witnessed 2 to 3 ots tht disregrded the chnnel signs nd illeglly used their comustion engines to trnsverse regulted segrss flts, potentilly creting new scrs. Some of these scrs were in the re of my experiments. Of the 70 experimentl scrs nd the 10 controls in my experiment, 14 were isected y new propeller scr over the course of the yer (Figure 44). Some oters severed pvc mrker poles with their min engine motor propeller (Figure 45). Six of 160 poles were dmged due to ot trffic over one yer. Bot ctivity nd ioturtion of scrs from rys were the most common disturnces to scrs nd were likely the key erosionl forces during the experiment. 30

43 Discussion Nturl recovery By fr the most successful wy to restore ll types of scrs is to protect segrss from ot trffic nd let it nturlly recover. I predicted tht different severities in individul propeller scrs would cuse different recovery times for H. wrightii, nd different recovery times were oserved. In the nturl recovery experiment, the scrs tht hd the highest recovery within yer were those tht strted with intct H. wrightii rhizomes nd cut ldes (Type 4 scrs) (Figure 9). The Type 4 scrs grew to the control H. wrightii density within 2 months nd were equl to the control in ll segrss nd scr mesurements fter 3 months. Types 1, 2, nd 3 did not recover y the end of 1 yer to control densities of percent cover, percent root cover, cnopy height, scr width, slope ngle, nd depth. Since there ws depth nd width remining in scr Types 1-3, H. wrightii coloniztion ws reduced. Recovery of Type 1 scrs which hd depth ws not significntly different in segrss mesurements from tht of Type 2 nd 3, which did not hve depth, showing tht depth of 9 cm does not slow segrss recoloniztion in scrs in Mosquito Lgoon. The lck of difference etween segrss recoloniztion of Type 1 scrs, nd Types 2 nd 3 mye due to lef litter cting s nturl fill in scrs hving depth. If the current rte of recovery continues in ech mesured vrile, the men estimted recovery time for H. wrightii is 25 months for Type 1, nd 19 months in Types 2 nd 3. Segrss restortion Segrss restortion is chllenging endevor, nd few studies hve hd success in estlishing new segrss growth from trnsplnts (Fonsec et l. 1998). Low survivl nd filure of segrss trnsplnts is ll too common (Fonsec et l. 1998). Fonsec et l. (1998) documented tht the men trnsplnt survivl ws 47% out of 53 pulished restortion studies. Cmpell 31

44 (2002) documented tht less thn 22% of pulished trnsplnttion efforts in Austrli were successful in chieving 50% trnsplnt survivl. However, Hll et l. (2006) showed tht the stple method ws successful in trnsplnting H. wrightii into scrs in Tmp By nd the Florid Keys, where trnsplnts estlished 100% H. wrightii cover fter 1 yer in oth loctions. In my experiment, I tested comintion of two previously successful nd cost-effective restortion methods used y Hll et l. (2006): filling the scr with snd, nd plcing trnsplnts of H. wrightii in the scr using grden stples. I predicted tht the ddition of different restortion methods would cuse Type 1 scrs to recover fster thn nturl recovery, however none of the restortion methods enhnced the recovery time. Mny of my scrs filled with snd were eroded over time. After 2 months, 25% lost ll fill snd nd fter one yer 75% of scrs lost ll fill snd completely (Tle 9). I hd hoped to hve high trnsplnt success, since (1) H. wrightii removed from the experiment ws replnted in the sme re, (2) trnsplnts were replnted within 1 to 4 hours the sme dy, nd (3) ll trnsplnt units of H. wrightii included growing tip t one end of the rhizome. However, ll trnsplnt units were lost within 2 months. Knowing tht stress to the trnsplnts ws minimized nd the control tretments showed no signs of die-off from wter qulity or sediment conditions, I hypothesize the min cuse of mortlity ws erosion, s evidenced y the men (± S.E.) sediment erosion depth of 7 (±3) cm under stples. Bot ctivity nd ioturtion of scrs from rys were the most common disturnces to scrs nd were likely the key erosionl forces during the experiment. Despite erosion forces, the mjority of the 70 experimentl scrs exhiited H. wrightii recoloniztion during the yer. Eight scrs (10%), recovered to 100% H. wrightii segrss cover within 1 yer. Nineteen scrs (20%) recovered to greter thn 75% cover. Twenty-nine scrs (40%) recovered to greter thn 50% cover. Only four scrs (5%) did not show ny recoloniztion 32

45 or recovery inside the scr within 1 yer (Figures 26, 34). Of those four scrs, three were fill nd plnt tretments nd one ws Type 2 scr. My experiment hs provided helpful informtion for resource mngers. First, the stple method ws not productive trnsplnt method for H. wrightii in Mosquito Lgoon. Only one trnsplnt out of 320 deployed in the experiment survived t 2 months. Second, lthough the sndfilled scrs provided solid sustrte on which segrss could colonize, there were no significnt differences in ny growth mesurements etween filled tretments nd the nturl recovery of Type 1 scrs. Thus, using snd to fill trenches did not enhnce recovery nd ws not n effective restortion method for H. wrightii in Mosquito Lgoon. Third, despite the slow currents of Mosquito Lgoon nd the protection of the segrss ed y Pole-Troll Zone regultions, erosion my hve plyed key role in why the restortion filed. Mngement recommendtions My results suggest tht if the Pole-Troll Zone regultions were strictly enforced, new scrs would not e creted nd ll current scrs would recover in 3 yers or less. If Pole-Troll Zone regultions were effective, then the mjority of the scrs in the segrss would e mde y trolling motors. Trolling motors hve djustle mounts for chnging propeller depth in shllow wter nd hve less power thn comustion engines which crete less severe scrs. Most trolling motors stop rotting if they rech the ottom, nd re not powerful enough to plow through the sediment nd impct segrss root structure s comustion engines cn. These less severe scrs (Type 4) could completely recover in 3 months or less. For res of segrss tht re more intensely scrred, creting temporry, no-motor zones tht prk mngement could rotte every 3 yers would llow enough time for the impcted 33

46 segrss eds to recover. Along with the recovery of segrss in no-motor zone, vrious species tht depend upon segrss would likely increse, such s commercilly importnt fish stock (Duffy 2006). More thn 119 species of fishes hve een documented in the IRL, nd segrss percent cover is n importnt fctor in determining fish ssemlges (Kupschus nd Tremin 2001). Mny fish species in the lgoon reproduce in the ocen, ut rely on esturine juvenile phse to complete the reproductive cycle (Kupschus nd Tremin 2001). For every segrss hectre in the IRL protected to llow recovery, or protected to prevention ot scrring, n dditionl $14,000 is gined in fisheries resources (IRL NEP 2008). Restortion versus prevention According to Engemn et l. (2008), segrss restortions re fr more expensive thn prevention methods tht involve dding signge nd lw enforcement officers. The dmge to segrss eds etween 1994 nd 2005 in the Florid Keys ws vlued t $28 million dollrs, nd estimted to increse $1 million in vlue ech consecutive yer (Engemn et l. 2008). However, this huge loss could hve een prevented with dditionl signs tht cost n verge of $192 ech, nd ptrol officers costing n verge of $59,400 per yer (Engemn et l. 2008). To put costs into perspective, single ptrol officer s nnul slry ws equivlent to the vlue of 0.42 h of segrss (Engemn et l. 2008). Previous segrss restortions in the Florid Keys hve een very costly t $940,000 per h to restore (Fonsec et l. 2002, Lewis et l. 2006). Besides the hefty price tg, segrss restortions cn tke over 10 yers to complete, nd there is no gurntee tht they will e successful (Engemn et l. 2008). For exmple, segrss mitigtion project for renovtion of Port Mntee in Tmp By compred mny segrss restortion techniques, few of which were successful (Lewis et l. 2006). The cost to trnsplnt 2.94 h of segrss nd restore 1.86 h of segrss ws $3.3 million per h of segrss restored (Lewis et l. 2006). The 2.94 h of 34

47 segrss ws mitigted s trnsplnts to enhnce res of degrded segrss from ot trffic (Lewis et l. 2006). The restortion methods of hnd-trnsplnting individul shoots, hndplnting shoot undles, hnd-plnting undles in pet pots, mechnized trnsplnting, modified mnul shovel plnting, nd prevention methods of creting no motor zone comined were successful in estlishing nd preserving 1.86 h of segrss (Lewis et l. 2006). The lest costly nd most successful method used to restore segrss eds ws to crete no motor zone in which 0.77 h of segrss recolonized propeller scrred res in 18 months (Lewis et l. 2006). Future of segrss in Mosquito Lgoon Segrss ecosystems re criticl in creting hitt in res of rren sediment, enhncing community diversity, estlishing intricte food wes, nd stilizing costl sediments (Beck et l. 2001, Burfeind nd Stunz 2006, Duffy 2006). Segrss eds in the Indin River Lgoon provide criticl hitt for over 2000 species of invertertes, fishes, irds nd mmmls, nd re key reson why the IRL is the most productive nd species-rich esturies in North Americ (IRL NEP 2008, Smithsonin Institution 2007). In my study, I looked t mjor thret to segrss ecosystems, ot propeller scrring. I documented for the first time the severity of ot propeller scrring in the IRL. In my mnipultive experiments, I found tht nturl recovery ws fster thn ny restortion ttempted method. Nturl recovery of propeller scrs cn e s quick s 3 months if some of the root structure nd ldes remin intct. Though my restortion methods were not effective for scrs in Mosquito Lgoon, the scrs were nturlly recolonized. Erosion ws prolem in oth experiments, ut despite ioturtion from rys nd dditionl ot scrring, 40% of ll experimentl scrs returned to greter thn 50% H. wrightii cover fter 1 yer. I gree with Engemn et l. (2008) tht more resources should focus on the prevention of scrring segrss eds thn which restortion methods would work est, especilly for Mosquito Lgoon where there re 35

48 thousnds of hectres of helthy segrss eds to recolonize scrred res. Enhnced vigilnce is required from users, mngers, nd enforcers to protect the segrss eds tht re criticl to the iodiversity of Mosquito Lgoon nd the IRL. Surely, prevention is the est wy to preserve our segrss eds nd the ssocited iodiversity for the future. 36

49 APPENDIX A: LIST OF FIGURES 37

50 Figure 1: Mp of experimentl scr loctions in the Pole-Troll Zone of Mosquito Lgoon, Florid. Boundry of the Pole-Troll Zone is mrked y dotted line with lck circles. Chnnels include: the Atlntic Intercostl Wterwy (ICW) s solid lck line, nd the inner Pole-Troll Zone chnnels re dotted lines within the Pole-Troll Zone oundries. 38

51 Figure 2: Hlodule wrightii segrss ed in Mosquito Lgoon, FL. 39

52 Figure 3: Loction of 110 propeller scrs mesured in Mosquito Lgoon, FL. Scr type is color coded to show the frequency of ech scr type. From this view, the entire rnge of mesured scrs is demonstrted, however not ll scrs cn e seen due to overlp. 40

53 Function 2 Segrss undnce Scr dimensions Function 1 Figure 4: Cnonicl Discriminnt Anlysis clssifiction of 110 propeller scrs in Mosquito Lgoon. Scrs were sorted into four groups. Type 1 scrs re white circles, Type 2 scrs re lck squres, Type 3 scrs re lck tringles, nd Type 4 scrs re white dimonds. 41

54 Figure 5: Scr types identified from mesuring 110 scrs in Mosquito Lgoon. Aove digrm simplifies wht ws seen in ech of the photogrphs elow. 27.3% 2 6.4% % 10.0% Figure 6: Frequency of scr types in Mosquito Lgoon (n = 110). 42

55 Scr Type 1 depth (cm) Scr Type 1 depth (cm) Scr Scr Type Type 1 width width (cm) (cm) Figure 7: Regression of Type 1 scr width versus scr depth. Scr Type 1 ws used to show the est correltion (r 2 = 0.476, p < 0.001), ecuse other scr types (2-4) did not hve depth. 43

56 Men % frction (g) Scr Type Scr Type Figure 8: Men percent of sediment grin size frctions nd orgnic content per scr type. Percent sediment frction sizes re listed in order < 1mm is light grey cross htch, 1 mm is drk grey cross htch, 2 mm is solid lck, 5 mm is solid white, > 5mm is solid light grey, nd percent orgnic content is solid drk grey. Using one-wy ANOVA, only scr Type 1 nd 4 were significntly different in ll frction sizes, nd size < 1mm showed the gretest difference. 44

57 sediment type Sediment type Scr distnce to chnnel scr distnce to chnnel 2 Figure 9: Cnonicl Discriminnt Anlysis clssifiction of 110 propeller scrs sediment frctions nd distnce mesurements. Type 1 scrs re white circles, Type 2 scrs re lck squres, Type 3 scrs re lck tringles, nd Type 4 scrs re white dimonds. Sediment types did not show pttern in reltion to distnce. 45

58 Figure 10: Lyout of one lock of experimentl scrs. Ech lock ws 18 x18 m nd ws divided into 9 squres, of which only 8 hd scr tretment. A squre ws 6 x 6 m contining single scr tretment in the center. 6 m Figure 11: Setup of one set of tretments in lrge grid (lock). An 18 x 18 m perimeter ws mrked off with yellow rope, nd PVC pole ws plced every 6 m. Scrs were creted in the center of ech 6 x 6 m squre nd mrked with short PVC pole t ech end. All PVC grid poles were mrked with GPS loction nd then removed. After the scrs were creted, only the 2 PVC scr mrkers remined to mrk the ends of ech scr. 46

59 2.5 cm Figure 12: PVC stke mrking the end of n experimentl scr. 32 cm Figure 13: PVC plow 32 cm in dimeter. Device used to mke experimentl propeller scrs the sme width, depth, nd scr slope ngle. 47

60 4 cm Figure 14: Arrow points to H. wrightii ttched to grden stple with twist tie in scr filled with snd. This is n eroded stple, mking it is esier to see how H. wrightii ws ttched with red twist tie. 48

61 T e m p e r tu re (C ) A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 15: Men wter temperture on smpling dtes (± S.E.) t loction of mnipultive experiments (n = 3) S lin ity (p p t) A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 16: Men slinity (± S.E.) t loction of mnipultive experiments on smpling dtes (n = 3). 49

62 60 55 w te r c l rity d e p th (c m ) A u g -0 7 O c t-0 7 D e c -0 7 F e -0 8 A p r-0 8 J u n -0 8 A u g -0 8 Figure 17: Men wter clrity depth (± S.E.) t loction of mnipultive experiments on smpling dtes (n = 3). Mximum mesurement possile using turitdity tue ws 60 cm d e p th (c m ) A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 18: Men wter depth (± S.E.) of experimentl scr loctions of mnipultive experiments (n = 10). Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 50

63 H lodule shoots A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 19: Men surrounding scr H. wrightii density (± S.E.) per 25 x 25 cm re in experimentl scr replicte locks. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters % H lo d u le cove r A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 20: Men surrounding scr H. wrightii percent cover (± S.E.) per 25 x 25 cm re in experimentl scr replicte locks. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 51

64 cnopy height (cm ) A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 21: Men surrounding scr H. wrightii cnopy height (± S.E.) in experimentl scr replicte locks. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 52

65 H lo d u le s h o o ts c 0 A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 22: Nturl recovery experimentl scr H. wrightii shoot density per 25 x 25 cm re (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, Type 4 re white dimonds, nd control re lck tringles. Mesurements were nlyzed with Kruskl- Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 53

66 % H lo d u le c o ve r c A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 23: Nturl recovery experimentl scr H. wrightii percent cover per 25 x 25 cm re (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, Type 4 re white dimonds, nd control re lck tringles. Mesurements were nlyzed with Kruskl- Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 54

67 50 40 cnopy h e ig h t (cm ) c c c 10 d c 0 c d A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 24: Nturl recovery experimentl scr H. wrightii cnopy height (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, Type 4 re white dimonds, nd control re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 55

68 % ro o t c o ve r c 0 A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 25: Nturl recovery experimentl scr percent root cover per 25 x 25 cm re (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, nd Type 4 re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 56

69 % le f litte r cove r A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 26: Nturl recovery experimentl scr percent lef litter cover per 25 x 25 cm re (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, nd Type 4 re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 57

70 12 10 s c r d e p th (c m ) c c c c c A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 27: Nturl recovery experimentl scr depth (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, nd Type 4 re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 58

71 60 50 s c r w id th (c m ) A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 28: Nturl recovery experimentl scr width (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, nd Type 4 re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 59

72 50 40 scr n g le in d e g re e s c A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 29: Nturl recovery experimentl scr slope ngle density (± S.E). Type 1 re lck circles, Type 2 re white tringles, Type 3 re lck squres, nd Type 4 re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 60

73 60 P e rc e n t fr c tio n In iti l F in l: T yp e 1 F in l: T yp e 2 F in l: T yp e 3 F in l: T yp e 4 F in l: C o n tro l 10 0 > 5 m m 5 m m 2 m m 1 m m < 1 m m Figure 30: Nturl recovery experiment men percent sediment frctions (± S.E.) from efore scr cretion nd from within scrs fter 1 yer (Kruskl-Wllis: p < 0.05). Br sections re leled in order from left to right: initil, Type 1 (finl), Type 2 (finl), Type 3 (finl), Type 4 (finl), nd control (finl). 61

74 6 5 P e rc e n t o rg n ic In iti l F in l: T yp e 1 F in l: T yp e 2 F in l: T yp e 3 F in l: T yp e 4 F in l: C o n tro l Figure 31: Nturl recovery experiment men percent orgnic content (± S.E) from efore scr cretion nd from within scrs fter 1 yer (Kruskl-Wllis: p < 0.05). 62

75 H lo d u le shoots A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 32: Restortion experimentl scr H. wrightii density shoot density per 25 x 25 cm (± S.E). Controls re white tringles, Type 1 re lck circles, plnt-only re lck squres, fill only re white dimonds, nd plnt nd fill re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 63

76 % H lo d u le c o ve r A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 33: Restortion experimentl scr H. wrightii percent cover per 25 x 25 cm re (± S.E). Control re white tringles, Type 1 re lck circles, plnt-only re lck squres, fill only re white dimonds, nd plnt nd fill re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 64

77 50 40 cnopy h e ig h t (cm ) A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 34: Restortion experimentl scr H. wrightii cnopy height (± S.E). Control re white tringles, Type 1 re lck circles, plnt-only re lck squres, fill only re white dimonds, nd plnt nd fill re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 65

78 60 50 % ro o t c o ve r A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 35: Restortion experimentl scr percent root cover per 25 x 25 cm re (± S.E). Type 1 re lck circles, plnt-only re white tringles, fill only re lck squres, nd plnt nd fill re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 66

79 % le f litte r cove r A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 36: Restortion experimentl scr percent lef litter cover per 25 x 25 cm re (± S.E). Type 1 re lck circles, plnt-only re white tringles, fill only re lck squres, nd plnt nd fill re white dimonds. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 67

80 s c r d e p th (c m ) A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 37: Restortion experimentl scr depth (± S.E). Control re white tringles, Type 1 re lck circles, plnt-only re lck squres, fill only re white dimonds, nd plnt nd fill re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 68

81 50 40 s c r w id th (c m ) A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 38: Restortion experimentl scr width (± S.E). Control re white tringles, Type 1 re lck circles, plnt-only re lck squres, fill only re white dimonds, nd plnt nd fill re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 69

82 50 40 s c r n g le in d e g re e s A u g 0 7 O c t0 7 D e c 0 7 F e 0 8 A p r0 8 J u n 0 8 A u g 0 8 Figure 39: Restortion experimentl scr slope ngle density (± S.E). Control re white tringles, Type 1 re lck circles, plnt-only re lck squres, fill only re white dimonds, nd plnt nd fill re lck tringles. Mesurements were nlyzed with Kruskl-Wllis rnk test with Bonferroni correction of p < Tretments tht were significntly different on single dy re shown with different letters. 70

83 60 P e rc e n t fr c tio n In iti l F in l: C o n tro l F in l: T yp e 1 F in l: P l n t o n ly F in l: F ill o n ly F in l: P l n t & fill 10 0 > 5 m m 5 m m 2 m m 1 m m < 1 m m Figure 40: Restortion experiment men percent sediment frctions (± S.E.) from efore scr cretion nd from within scrs fter 1 yer (Kruskl-Wllis: p < 0.05). Br sections re leled in order from left to right: initil, control (finl), Type 1 (finl), plnt only (finl), fill only (finl), nd plnt nd fill (finl). 71

84 6 5 P e rc e n t o rg n ic In iti l F in l: C o n tro l F in l: P l n t o n ly F in l: F ill o n ly F in l: P l n t & fill Figure 41: Restortion experiment men percent orgnic content (± S.E) from efore scr cretion nd from within scrs fter 1 yer (Kruskl-Wllis: p < 0.05). 72

85 Figure 42: Types of erosion oserved in experimentl scrs: A) holes, B) wshouts, nd C) snd ptches. The lck rectngle represents the originl scr dimensions, rown is re sediment, drk rown represents depth, nd the green represents segrss. 4 cm 4 cm Figure 43: Bioturtion cused y stingry. Stingry creted hole nd unuried row of Hlodule wrightii trnsplnts on stples, then left ehind pile of feces. In the left picture the rrow points to close-up of excrement pile in hole, nd in right picture the rrow points to close up of the eroded stples next to the excrement pile. 73

86 Figure 44: Boter illeglly using min engine motor in Pole-Troll Zone nd running directly over the experiment site. 2 cm Figure 45: A PVC pole mrking n experimentl scr tht ws hit y min engine ot motor in less thn 0.5 m of wter. 74